vertex ai agent engine terraform provider

[jsteinberg-rbi]

Is your feature request related to a problem? Please describe.
First, I just want to say how much I appreciate the quality of the ADK and Vertex AI Agent Engine documentation — it’s truly excellent. It’s been very easy to get started, and I’m grateful for how thoughtfully it’s been put together.
As I explore scaling up, I’m imagining a scenario where 30 teams each deploy 3 agentic services, with each service hosting 3 agents. That would result in 270 agents being instantiated programmatically. Today, it seems there isn’t a straightforward way to manage their lifecycle — particularly for destroying them — without either custom-building my own logic or manually deleting them one by one.
I noticed that other parts of Vertex AI already have Terraform provider resources, and I would like to suggest that creating Terraform support for Agent Engine could become very important. As organizations encourage agentic development at scale, it feels crucial that Google provides a first-class way to orchestrate agents — not just develop and host them.
While Terraform may not always be ideal for application-level deployment, it remains the standard for managing the infrastructure that supports applications, especially for teams that aren't using Kubernetes (where options like Operators, Helm, etc., are more common). Even though Agent Engine is serverless, teams will still need a way to programmatically create and configure reasoning engine instances reliably.

Describe the solution you'd like
A suite of Terraform provider resources that fully support the Agent Engine lifecycle.

Describe alternatives you've considered

1. Each team independently building their own create/destroy workflows
2. Manual deletion of agents, which could become laborious and error-prone over time

[iker-gonzalez]

This would be awesome

[wdai-chwy]

+1. right now, agent engine has to be deployed by python script or CLI. this is not the best way to manage infra resource.